1. Field of the Invention
The present invention relates to forming features of an integrated circuit (IC) and in particular achieving sub-wavelength resolution of certain features in the IC in a cost-effective manner.
2. Description of the Related Art
In designing an integrated circuit (IC), engineers typically rely upon computer simulation tools to help create a schematic circuit design consisting of individual circuit elements electrically coupled together in order to perform certain functions. To actually fabricate this integrated circuit in a semiconductor substrate, the circuit design must be translated into a physical representation, or layout, which itself can then be transferred onto a series of templates (e.g. masks) used to sequentially pattern layers in or on the semiconductor substrate surface. Computer aided design (CAD) tools assist layout designers in the task of translating the circuit design into a series of two-dimensional patterns that will define the component layers of the IC, such as the active device regions, gate electrodes, contact holes, metal interconnections, and so on.
A method of transferring a layout pattern to the semiconductor substrate surface is to use the process of optical lithography (photolithography) in which the layout pattern is first transferred onto a physical template that in turn is used to optically project the layout pattern onto the surface of the semiconductor substrate (hereinafter wafer).
In transferring the IC layout to physical templates, a mask is generally created for each layer of the IC. For example, the data representing the layout pattern for a specific layer (e.g. the gate electrode layer) can be input into an electron beam machine that writes the layout pattern onto a blank mask. After the mask is made, it is used to optically project the layout pattern onto many wafers, one at a time. This optical projection is done by shining light through the mask onto the wafer. Optical lenses and/or mirrors may be used to direct, demagnify, and/or focus the mask image onto the wafer surface. Prior to exposure, the wafer is first coated with a masking layer of photosensitive material that is resistant to being etched and is hence referred to as photoresist.
For a binary mask, light passes through the clear regions of the mask, thereby exposing the photoresist coating in these regions. In contrast, light is blocked by the opaque regions of the binary mask, thereby leaving the photoresist coating unexposed in these regions. When the photoresist coating is then developed in a chemical solution, either the exposed regions (for a positive photoresist) or unexposed regions (for a negative photoresist) are selectively removed. The end result is a wafer coated with a layer of photoresist exhibiting a desired pattern to define the geometries, features, lines, and shapes of an underlying layer or an overlying layer. The photoresist layer is then removed after the underlying layer is processed (e.g. etched) or after the overlying layer is deposited, respectively. This photolithography process is used to define each layer of the IC, generally using a separate mask for each layer.
FIG. 1 illustrates a graph 100 plotting length (on a logarithmic scale) versus year. As indicated, the wavelength of light used in photolithography (shown by curve 101) to define features on a wafer was shorter than the minimum lithographically defined feature size of an IC (shown by curve 102) before 1996, i.e. until approximately the 0.25 μm (minimum half-pitch) technology node. In this time period, synthesis of the layout patterns and their transfer from the mask to the wafer were relatively straightforward with minimal distortions. For example, FIG. 2 illustrates features 204, 205, and 206 at the 0.25 μm (250 nm) technology node, which were generated during a design stage 201, a mask stage 202, and a wafer stage 203, respectively. At this technology node, a mask can merely comprise the geometric shapes that represent the desired layout pattern for its corresponding layer.
As indicated in graph 100 (FIG. 1), after the 0.25 μm technology node, the minimum feature size has been increasingly smaller than the wavelength of light used in photolithography. Thus, in many CMOS (complementary metal-oxide-semiconductor) IC products currently manufactured, the minimum feature size (e.g. the minimum gate length Lgmin of a transistor) is much smaller than the wavelength of light used in the photolithography process to define it. In this sub-wavelength photolithography regime, resolution enhancement techniques (RETs) are required at mask stage 202 to achieve the desired layout patterns on the wafer, i.e. at wafer stage 203.
For example, at the 0.18 μm (180 nm) technology node, shown in FIG. 2, a design feature 207, if merely reproduced as mask feature 208, would result in a poorly defined wafer feature 210. To achieve acceptable definition, RETs such as rule-based optical proximity correction (OPC) and model-based OPC can be used to generate an OPC-corrected mask feature 209, which in turn can be used generate a wafer feature 211. Rule-based OPC features can include serifs, hammerheads, and assist bars. In model-based OPC, edge segments of the design feature can be moved. In either OPC approach, the original design feature is modified to compensate for proximity effects.
Smaller technology nodes require yet more complex layout patterns at the mask stage. For example, at the 0.09 μm (90 nm) technology node and beyond, a design feature 212 reproduced simply as mask feature 213 will not even print in wafer stage 203. Another RET, called phase shifting, can be used to generate a phase-shifted mask feature 214. Phase shifting, in general, enhances the contrast of the lithography process. However, at this technology node, phase shifting by itself can only generate a poorly defined wafer feature 216. Therefore, a combination of OPC and phase shifting RETs is needed to generate feature 215, which in turn can generate a wafer feature 217 that is true to the design feature 212.
Notably, such complex RETs can make sub-wavelength resolution possible, but at increased design and manufacturing (e.g. photolithography) cost. Unfortunately, because of the widening gap between the wavelength of light and the minimum feature size with technology advancement over time, this cost is expected to significantly increase with each new generation of integrated circuit technology.
Therefore a need arises for a technique to provide good sub-wavelength feature definition in a cost effective manner.